add json file and txt file

added_tokens.json

@@ -0,0 +1,4 @@
+{
+  "<|im_end|>": 100279,
+  "<|im_start|>": 100278
+}

generation_config.json

@@ -0,0 +1,8 @@
+{
+  "_from_model_config": true,
+  "eos_token_id": [
+    100257,
+    100279
+  ],
+  "transformers_version": "4.45.0"
+}

model.safetensors.index.json

@@ -0,0 +1,770 @@
+{
+  "metadata": {
+    "total_size": 132834094880
+  },
+  "weight_map": {
+    "lm_head.weight": "model-00031-of-00031.safetensors",
+    "transformer.blocks.0.ffn.experts.mlp.v1.input_scale": "model-00001-of-00031.safetensors",
+    "transformer.blocks.0.ffn.experts.mlp.v1.weight": "model-00001-of-00031.safetensors",
+    "transformer.blocks.0.ffn.experts.mlp.v1.weight_scale": "model-00001-of-00031.safetensors",
+    "transformer.blocks.0.ffn.experts.mlp.w1.input_scale": "model-00001-of-00031.safetensors",
+    "transformer.blocks.0.ffn.experts.mlp.w1.weight": "model-00001-of-00031.safetensors",
+    "transformer.blocks.0.ffn.experts.mlp.w1.weight_scale": "model-00001-of-00031.safetensors",
+    "transformer.blocks.0.ffn.experts.mlp.w2.input_scale": "model-00001-of-00031.safetensors",
+    "transformer.blocks.0.ffn.experts.mlp.w2.weight": "model-00001-of-00031.safetensors",
+    "transformer.blocks.0.ffn.experts.mlp.w2.weight_scale": "model-00001-of-00031.safetensors",
+    "transformer.blocks.0.ffn.router.layer.weight": "model-00001-of-00031.safetensors",
+    "transformer.blocks.0.norm_attn_norm.attn.Wqkv.input_scale": "model-00001-of-00031.safetensors",
+    "transformer.blocks.0.norm_attn_norm.attn.Wqkv.weight": "model-00001-of-00031.safetensors",
+    "transformer.blocks.0.norm_attn_norm.attn.Wqkv.weight_scale": "model-00001-of-00031.safetensors",
+    "transformer.blocks.0.norm_attn_norm.attn.kv_scale": "model-00001-of-00031.safetensors",
+    "transformer.blocks.0.norm_attn_norm.attn.out_proj.input_scale": "model-00001-of-00031.safetensors",
+    "transformer.blocks.0.norm_attn_norm.attn.out_proj.weight": "model-00001-of-00031.safetensors",
+    "transformer.blocks.0.norm_attn_norm.attn.out_proj.weight_scale": "model-00001-of-00031.safetensors",
+    "transformer.blocks.0.norm_attn_norm.norm_1.weight": "model-00001-of-00031.safetensors",
+    "transformer.blocks.0.norm_attn_norm.norm_2.weight": "model-00001-of-00031.safetensors",
+    "transformer.blocks.1.ffn.experts.mlp.v1.input_scale": "model-00002-of-00031.safetensors",
+    "transformer.blocks.1.ffn.experts.mlp.v1.weight": "model-00002-of-00031.safetensors",
+    "transformer.blocks.1.ffn.experts.mlp.v1.weight_scale": "model-00002-of-00031.safetensors",
+    "transformer.blocks.1.ffn.experts.mlp.w1.input_scale": "model-00002-of-00031.safetensors",
+    "transformer.blocks.1.ffn.experts.mlp.w1.weight": "model-00002-of-00031.safetensors",
+    "transformer.blocks.1.ffn.experts.mlp.w1.weight_scale": "model-00002-of-00031.safetensors",
+    "transformer.blocks.1.ffn.experts.mlp.w2.input_scale": "model-00002-of-00031.safetensors",
+    "transformer.blocks.1.ffn.experts.mlp.w2.weight": "model-00002-of-00031.safetensors",
+    "transformer.blocks.1.ffn.experts.mlp.w2.weight_scale": "model-00002-of-00031.safetensors",
+    "transformer.blocks.1.ffn.router.layer.weight": "model-00001-of-00031.safetensors",
+    "transformer.blocks.1.norm_attn_norm.attn.Wqkv.input_scale": "model-00001-of-00031.safetensors",
+    "transformer.blocks.1.norm_attn_norm.attn.Wqkv.weight": "model-00001-of-00031.safetensors",
+    "transformer.blocks.1.norm_attn_norm.attn.Wqkv.weight_scale": "model-00001-of-00031.safetensors",
+    "transformer.blocks.1.norm_attn_norm.attn.kv_scale": "model-00001-of-00031.safetensors",
+    "transformer.blocks.1.norm_attn_norm.attn.out_proj.input_scale": "model-00001-of-00031.safetensors",
+    "transformer.blocks.1.norm_attn_norm.attn.out_proj.weight": "model-00001-of-00031.safetensors",
+    "transformer.blocks.1.norm_attn_norm.attn.out_proj.weight_scale": "model-00001-of-00031.safetensors",
+    "transformer.blocks.1.norm_attn_norm.norm_1.weight": "model-00001-of-00031.safetensors",
+    "transformer.blocks.1.norm_attn_norm.norm_2.weight": "model-00001-of-00031.safetensors",
+    "transformer.blocks.10.ffn.experts.mlp.v1.input_scale": "model-00009-of-00031.safetensors",
+    "transformer.blocks.10.ffn.experts.mlp.v1.weight": "model-00009-of-00031.safetensors",
+    "transformer.blocks.10.ffn.experts.mlp.v1.weight_scale": "model-00009-of-00031.safetensors",
+    "transformer.blocks.10.ffn.experts.mlp.w1.input_scale": "model-00008-of-00031.safetensors",
+    "transformer.blocks.10.ffn.experts.mlp.w1.weight": "model-00008-of-00031.safetensors",
+    "transformer.blocks.10.ffn.experts.mlp.w1.weight_scale": "model-00008-of-00031.safetensors",
+    "transformer.blocks.10.ffn.experts.mlp.w2.input_scale": "model-00009-of-00031.safetensors",
+    "transformer.blocks.10.ffn.experts.mlp.w2.weight": "model-00009-of-00031.safetensors",
+    "transformer.blocks.10.ffn.experts.mlp.w2.weight_scale": "model-00009-of-00031.safetensors",
+    "transformer.blocks.10.ffn.router.layer.weight": "model-00008-of-00031.safetensors",
+    "transformer.blocks.10.norm_attn_norm.attn.Wqkv.input_scale": "model-00008-of-00031.safetensors",
+    "transformer.blocks.10.norm_attn_norm.attn.Wqkv.weight": "model-00008-of-00031.safetensors",
+    "transformer.blocks.10.norm_attn_norm.attn.Wqkv.weight_scale": "model-00008-of-00031.safetensors",
+    "transformer.blocks.10.norm_attn_norm.attn.kv_scale": "model-00008-of-00031.safetensors",
+    "transformer.blocks.10.norm_attn_norm.attn.out_proj.input_scale": "model-00008-of-00031.safetensors",
+    "transformer.blocks.10.norm_attn_norm.attn.out_proj.weight": "model-00008-of-00031.safetensors",
+    "transformer.blocks.10.norm_attn_norm.attn.out_proj.weight_scale": "model-00008-of-00031.safetensors",
+    "transformer.blocks.10.norm_attn_norm.norm_1.weight": "model-00008-of-00031.safetensors",
+    "transformer.blocks.10.norm_attn_norm.norm_2.weight": "model-00008-of-00031.safetensors",
+    "transformer.blocks.11.ffn.experts.mlp.v1.input_scale": "model-00009-of-00031.safetensors",
+    "transformer.blocks.11.ffn.experts.mlp.v1.weight": "model-00009-of-00031.safetensors",
+    "transformer.blocks.11.ffn.experts.mlp.v1.weight_scale": "model-00009-of-00031.safetensors",
+    "transformer.blocks.11.ffn.experts.mlp.w1.input_scale": "model-00009-of-00031.safetensors",
+    "transformer.blocks.11.ffn.experts.mlp.w1.weight": "model-00009-of-00031.safetensors",
+    "transformer.blocks.11.ffn.experts.mlp.w1.weight_scale": "model-00009-of-00031.safetensors",
+    "transformer.blocks.11.ffn.experts.mlp.w2.input_scale": "model-00010-of-00031.safetensors",
+    "transformer.blocks.11.ffn.experts.mlp.w2.weight": "model-00010-of-00031.safetensors",
+    "transformer.blocks.11.ffn.experts.mlp.w2.weight_scale": "model-00010-of-00031.safetensors",
+    "transformer.blocks.11.ffn.router.layer.weight": "model-00009-of-00031.safetensors",
+    "transformer.blocks.11.norm_attn_norm.attn.Wqkv.input_scale": "model-00009-of-00031.safetensors",
+    "transformer.blocks.11.norm_attn_norm.attn.Wqkv.weight": "model-00009-of-00031.safetensors",
+    "transformer.blocks.11.norm_attn_norm.attn.Wqkv.weight_scale": "model-00009-of-00031.safetensors",
+    "transformer.blocks.11.norm_attn_norm.attn.kv_scale": "model-00009-of-00031.safetensors",
+    "transformer.blocks.11.norm_attn_norm.attn.out_proj.input_scale": "model-00009-of-00031.safetensors",
+    "transformer.blocks.11.norm_attn_norm.attn.out_proj.weight": "model-00009-of-00031.safetensors",
+    "transformer.blocks.11.norm_attn_norm.attn.out_proj.weight_scale": "model-00009-of-00031.safetensors",
+    "transformer.blocks.11.norm_attn_norm.norm_1.weight": "model-00009-of-00031.safetensors",
+    "transformer.blocks.11.norm_attn_norm.norm_2.weight": "model-00009-of-00031.safetensors",
+    "transformer.blocks.12.ffn.experts.mlp.v1.input_scale": "model-00010-of-00031.safetensors",
+    "transformer.blocks.12.ffn.experts.mlp.v1.weight": "model-00010-of-00031.safetensors",
+    "transformer.blocks.12.ffn.experts.mlp.v1.weight_scale": "model-00010-of-00031.safetensors",
+    "transformer.blocks.12.ffn.experts.mlp.w1.input_scale": "model-00010-of-00031.safetensors",
+    "transformer.blocks.12.ffn.experts.mlp.w1.weight": "model-00010-of-00031.safetensors",
+    "transformer.blocks.12.ffn.experts.mlp.w1.weight_scale": "model-00010-of-00031.safetensors",
+    "transformer.blocks.12.ffn.experts.mlp.w2.input_scale": "model-00010-of-00031.safetensors",
+    "transformer.blocks.12.ffn.experts.mlp.w2.weight": "model-00010-of-00031.safetensors",
+    "transformer.blocks.12.ffn.experts.mlp.w2.weight_scale": "model-00010-of-00031.safetensors",
+    "transformer.blocks.12.ffn.router.layer.weight": "model-00010-of-00031.safetensors",
+    "transformer.blocks.12.norm_attn_norm.attn.Wqkv.input_scale": "model-00010-of-00031.safetensors",
+    "transformer.blocks.12.norm_attn_norm.attn.Wqkv.weight": "model-00010-of-00031.safetensors",
+    "transformer.blocks.12.norm_attn_norm.attn.Wqkv.weight_scale": "model-00010-of-00031.safetensors",
+    "transformer.blocks.12.norm_attn_norm.attn.kv_scale": "model-00010-of-00031.safetensors",
+    "transformer.blocks.12.norm_attn_norm.attn.out_proj.input_scale": "model-00010-of-00031.safetensors",
+    "transformer.blocks.12.norm_attn_norm.attn.out_proj.weight": "model-00010-of-00031.safetensors",
+    "transformer.blocks.12.norm_attn_norm.attn.out_proj.weight_scale": "model-00010-of-00031.safetensors",
+    "transformer.blocks.12.norm_attn_norm.norm_1.weight": "model-00010-of-00031.safetensors",
+    "transformer.blocks.12.norm_attn_norm.norm_2.weight": "model-00010-of-00031.safetensors",
+    "transformer.blocks.13.ffn.experts.mlp.v1.input_scale": "model-00011-of-00031.safetensors",
+    "transformer.blocks.13.ffn.experts.mlp.v1.weight": "model-00011-of-00031.safetensors",
+    "transformer.blocks.13.ffn.experts.mlp.v1.weight_scale": "model-00011-of-00031.safetensors",
+    "transformer.blocks.13.ffn.experts.mlp.w1.input_scale": "model-00011-of-00031.safetensors",
+    "transformer.blocks.13.ffn.experts.mlp.w1.weight": "model-00011-of-00031.safetensors",
+    "transformer.blocks.13.ffn.experts.mlp.w1.weight_scale": "model-00011-of-00031.safetensors",
+    "transformer.blocks.13.ffn.experts.mlp.w2.input_scale": "model-00011-of-00031.safetensors",
+    "transformer.blocks.13.ffn.experts.mlp.w2.weight": "model-00011-of-00031.safetensors",
+    "transformer.blocks.13.ffn.experts.mlp.w2.weight_scale": "model-00011-of-00031.safetensors",
+    "transformer.blocks.13.ffn.router.layer.weight": "model-00010-of-00031.safetensors",
+    "transformer.blocks.13.norm_attn_norm.attn.Wqkv.input_scale": "model-00010-of-00031.safetensors",
+    "transformer.blocks.13.norm_attn_norm.attn.Wqkv.weight": "model-00010-of-00031.safetensors",
+    "transformer.blocks.13.norm_attn_norm.attn.Wqkv.weight_scale": "model-00010-of-00031.safetensors",
+    "transformer.blocks.13.norm_attn_norm.attn.kv_scale": "model-00010-of-00031.safetensors",
+    "transformer.blocks.13.norm_attn_norm.attn.out_proj.input_scale": "model-00010-of-00031.safetensors",
+    "transformer.blocks.13.norm_attn_norm.attn.out_proj.weight": "model-00010-of-00031.safetensors",
+    "transformer.blocks.13.norm_attn_norm.attn.out_proj.weight_scale": "model-00010-of-00031.safetensors",
+    "transformer.blocks.13.norm_attn_norm.norm_1.weight": "model-00010-of-00031.safetensors",
+    "transformer.blocks.13.norm_attn_norm.norm_2.weight": "model-00010-of-00031.safetensors",
+    "transformer.blocks.14.ffn.experts.mlp.v1.input_scale": "model-00012-of-00031.safetensors",
+    "transformer.blocks.14.ffn.experts.mlp.v1.weight": "model-00012-of-00031.safetensors",
+    "transformer.blocks.14.ffn.experts.mlp.v1.weight_scale": "model-00012-of-00031.safetensors",
+    "transformer.blocks.14.ffn.experts.mlp.w1.input_scale": "model-00011-of-00031.safetensors",
+    "transformer.blocks.14.ffn.experts.mlp.w1.weight": "model-00011-of-00031.safetensors",
+    "transformer.blocks.14.ffn.experts.mlp.w1.weight_scale": "model-00011-of-00031.safetensors",
+    "transformer.blocks.14.ffn.experts.mlp.w2.input_scale": "model-00012-of-00031.safetensors",
+    "transformer.blocks.14.ffn.experts.mlp.w2.weight": "model-00012-of-00031.safetensors",
+    "transformer.blocks.14.ffn.experts.mlp.w2.weight_scale": "model-00012-of-00031.safetensors",
+    "transformer.blocks.14.ffn.router.layer.weight": "model-00011-of-00031.safetensors",
+    "transformer.blocks.14.norm_attn_norm.attn.Wqkv.input_scale": "model-00011-of-00031.safetensors",
+    "transformer.blocks.14.norm_attn_norm.attn.Wqkv.weight": "model-00011-of-00031.safetensors",
+    "transformer.blocks.14.norm_attn_norm.attn.Wqkv.weight_scale": "model-00011-of-00031.safetensors",
+    "transformer.blocks.14.norm_attn_norm.attn.kv_scale": "model-00011-of-00031.safetensors",
+    "transformer.blocks.14.norm_attn_norm.attn.out_proj.input_scale": "model-00011-of-00031.safetensors",
+    "transformer.blocks.14.norm_attn_norm.attn.out_proj.weight": "model-00011-of-00031.safetensors",
+    "transformer.blocks.14.norm_attn_norm.attn.out_proj.weight_scale": "model-00011-of-00031.safetensors",
+    "transformer.blocks.14.norm_attn_norm.norm_1.weight": "model-00011-of-00031.safetensors",
+    "transformer.blocks.14.norm_attn_norm.norm_2.weight": "model-00011-of-00031.safetensors",
+    "transformer.blocks.15.ffn.experts.mlp.v1.input_scale": "model-00012-of-00031.safetensors",
+    "transformer.blocks.15.ffn.experts.mlp.v1.weight": "model-00012-of-00031.safetensors",
+    "transformer.blocks.15.ffn.experts.mlp.v1.weight_scale": "model-00012-of-00031.safetensors",
+    "transformer.blocks.15.ffn.experts.mlp.w1.input_scale": "model-00012-of-00031.safetensors",
+    "transformer.blocks.15.ffn.experts.mlp.w1.weight": "model-00012-of-00031.safetensors",
+    "transformer.blocks.15.ffn.experts.mlp.w1.weight_scale": "model-00012-of-00031.safetensors",
+    "transformer.blocks.15.ffn.experts.mlp.w2.input_scale": "model-00013-of-00031.safetensors",
+    "transformer.blocks.15.ffn.experts.mlp.w2.weight": "model-00013-of-00031.safetensors",
+    "transformer.blocks.15.ffn.experts.mlp.w2.weight_scale": "model-00013-of-00031.safetensors",
+    "transformer.blocks.15.ffn.router.layer.weight": "model-00012-of-00031.safetensors",
+    "transformer.blocks.15.norm_attn_norm.attn.Wqkv.input_scale": "model-00012-of-00031.safetensors",
+    "transformer.blocks.15.norm_attn_norm.attn.Wqkv.weight": "model-00012-of-00031.safetensors",
+    "transformer.blocks.15.norm_attn_norm.attn.Wqkv.weight_scale": "model-00012-of-00031.safetensors",
+    "transformer.blocks.15.norm_attn_norm.attn.kv_scale": "model-00012-of-00031.safetensors",
+    "transformer.blocks.15.norm_attn_norm.attn.out_proj.input_scale": "model-00012-of-00031.safetensors",
+    "transformer.blocks.15.norm_attn_norm.attn.out_proj.weight": "model-00012-of-00031.safetensors",
+    "transformer.blocks.15.norm_attn_norm.attn.out_proj.weight_scale": "model-00012-of-00031.safetensors",
+    "transformer.blocks.15.norm_attn_norm.norm_1.weight": "model-00012-of-00031.safetensors",
+    "transformer.blocks.15.norm_attn_norm.norm_2.weight": "model-00012-of-00031.safetensors",
+    "transformer.blocks.16.ffn.experts.mlp.v1.input_scale": "model-00013-of-00031.safetensors",
+    "transformer.blocks.16.ffn.experts.mlp.v1.weight": "model-00013-of-00031.safetensors",
+    "transformer.blocks.16.ffn.experts.mlp.v1.weight_scale": "model-00013-of-00031.safetensors",
+    "transformer.blocks.16.ffn.experts.mlp.w1.input_scale": "model-00013-of-00031.safetensors",
+    "transformer.blocks.16.ffn.experts.mlp.w1.weight": "model-00013-of-00031.safetensors",
+    "transformer.blocks.16.ffn.experts.mlp.w1.weight_scale": "model-00013-of-00031.safetensors",
+    "transformer.blocks.16.ffn.experts.mlp.w2.input_scale": "model-00013-of-00031.safetensors",
+    "transformer.blocks.16.ffn.experts.mlp.w2.weight": "model-00013-of-00031.safetensors",
+    "transformer.blocks.16.ffn.experts.mlp.w2.weight_scale": "model-00013-of-00031.safetensors",
+    "transformer.blocks.16.ffn.router.layer.weight": "model-00013-of-00031.safetensors",
+    "transformer.blocks.16.norm_attn_norm.attn.Wqkv.input_scale": "model-00013-of-00031.safetensors",
+    "transformer.blocks.16.norm_attn_norm.attn.Wqkv.weight": "model-00013-of-00031.safetensors",
+    "transformer.blocks.16.norm_attn_norm.attn.Wqkv.weight_scale": "model-00013-of-00031.safetensors",
+    "transformer.blocks.16.norm_attn_norm.attn.kv_scale": "model-00013-of-00031.safetensors",
+    "transformer.blocks.16.norm_attn_norm.attn.out_proj.input_scale": "model-00013-of-00031.safetensors",
+    "transformer.blocks.16.norm_attn_norm.attn.out_proj.weight": "model-00013-of-00031.safetensors",
+    "transformer.blocks.16.norm_attn_norm.attn.out_proj.weight_scale": "model-00013-of-00031.safetensors",
+    "transformer.blocks.16.norm_attn_norm.norm_1.weight": "model-00013-of-00031.safetensors",
+    "transformer.blocks.16.norm_attn_norm.norm_2.weight": "model-00013-of-00031.safetensors",
+    "transformer.blocks.17.ffn.experts.mlp.v1.input_scale": "model-00014-of-00031.safetensors",
+    "transformer.blocks.17.ffn.experts.mlp.v1.weight": "model-00014-of-00031.safetensors",
+    "transformer.blocks.17.ffn.experts.mlp.v1.weight_scale": "model-00014-of-00031.safetensors",
+    "transformer.blocks.17.ffn.experts.mlp.w1.input_scale": "model-00014-of-00031.safetensors",
+    "transformer.blocks.17.ffn.experts.mlp.w1.weight": "model-00014-of-00031.safetensors",
+    "transformer.blocks.17.ffn.experts.mlp.w1.weight_scale": "model-00014-of-00031.safetensors",
+    "transformer.blocks.17.ffn.experts.mlp.w2.input_scale": "model-00014-of-00031.safetensors",
+    "transformer.blocks.17.ffn.experts.mlp.w2.weight": "model-00014-of-00031.safetensors",
+    "transformer.blocks.17.ffn.experts.mlp.w2.weight_scale": "model-00014-of-00031.safetensors",
+    "transformer.blocks.17.ffn.router.layer.weight": "model-00013-of-00031.safetensors",
+    "transformer.blocks.17.norm_attn_norm.attn.Wqkv.input_scale": "model-00013-of-00031.safetensors",
+    "transformer.blocks.17.norm_attn_norm.attn.Wqkv.weight": "model-00013-of-00031.safetensors",
+    "transformer.blocks.17.norm_attn_norm.attn.Wqkv.weight_scale": "model-00013-of-00031.safetensors",
+    "transformer.blocks.17.norm_attn_norm.attn.kv_scale": "model-00013-of-00031.safetensors",
+    "transformer.blocks.17.norm_attn_norm.attn.out_proj.input_scale": "model-00013-of-00031.safetensors",
+    "transformer.blocks.17.norm_attn_norm.attn.out_proj.weight": "model-00013-of-00031.safetensors",
+    "transformer.blocks.17.norm_attn_norm.attn.out_proj.weight_scale": "model-00013-of-00031.safetensors",
+    "transformer.blocks.17.norm_attn_norm.norm_1.weight": "model-00013-of-00031.safetensors",
+    "transformer.blocks.17.norm_attn_norm.norm_2.weight": "model-00013-of-00031.safetensors",
+    "transformer.blocks.18.ffn.experts.mlp.v1.input_scale": "model-00015-of-00031.safetensors",
+    "transformer.blocks.18.ffn.experts.mlp.v1.weight": "model-00015-of-00031.safetensors",
+    "transformer.blocks.18.ffn.experts.mlp.v1.weight_scale": "model-00015-of-00031.safetensors",
+    "transformer.blocks.18.ffn.experts.mlp.w1.input_scale": "model-00014-of-00031.safetensors",
+    "transformer.blocks.18.ffn.experts.mlp.w1.weight": "model-00014-of-00031.safetensors",
+    "transformer.blocks.18.ffn.experts.mlp.w1.weight_scale": "model-00014-of-00031.safetensors",
+    "transformer.blocks.18.ffn.experts.mlp.w2.input_scale": "model-00015-of-00031.safetensors",
+    "transformer.blocks.18.ffn.experts.mlp.w2.weight": "model-00015-of-00031.safetensors",
+    "transformer.blocks.18.ffn.experts.mlp.w2.weight_scale": "model-00015-of-00031.safetensors",
+    "transformer.blocks.18.ffn.router.layer.weight": "model-00014-of-00031.safetensors",
+    "transformer.blocks.18.norm_attn_norm.attn.Wqkv.input_scale": "model-00014-of-00031.safetensors",
+    "transformer.blocks.18.norm_attn_norm.attn.Wqkv.weight": "model-00014-of-00031.safetensors",
+    "transformer.blocks.18.norm_attn_norm.attn.Wqkv.weight_scale": "model-00014-of-00031.safetensors",
+    "transformer.blocks.18.norm_attn_norm.attn.kv_scale": "model-00014-of-00031.safetensors",
+    "transformer.blocks.18.norm_attn_norm.attn.out_proj.input_scale": "model-00014-of-00031.safetensors",
+    "transformer.blocks.18.norm_attn_norm.attn.out_proj.weight": "model-00014-of-00031.safetensors",
+    "transformer.blocks.18.norm_attn_norm.attn.out_proj.weight_scale": "model-00014-of-00031.safetensors",
+    "transformer.blocks.18.norm_attn_norm.norm_1.weight": "model-00014-of-00031.safetensors",
+    "transformer.blocks.18.norm_attn_norm.norm_2.weight": "model-00014-of-00031.safetensors",
+    "transformer.blocks.19.ffn.experts.mlp.v1.input_scale": "model-00015-of-00031.safetensors",
+    "transformer.blocks.19.ffn.experts.mlp.v1.weight": "model-00015-of-00031.safetensors",
+    "transformer.blocks.19.ffn.experts.mlp.v1.weight_scale": "model-00015-of-00031.safetensors",
+    "transformer.blocks.19.ffn.experts.mlp.w1.input_scale": "model-00015-of-00031.safetensors",
+    "transformer.blocks.19.ffn.experts.mlp.w1.weight": "model-00015-of-00031.safetensors",
+    "transformer.blocks.19.ffn.experts.mlp.w1.weight_scale": "model-00015-of-00031.safetensors",
+    "transformer.blocks.19.ffn.experts.mlp.w2.input_scale": "model-00016-of-00031.safetensors",
+    "transformer.blocks.19.ffn.experts.mlp.w2.weight": "model-00016-of-00031.safetensors",
+    "transformer.blocks.19.ffn.experts.mlp.w2.weight_scale": "model-00016-of-00031.safetensors",
+    "transformer.blocks.19.ffn.router.layer.weight": "model-00015-of-00031.safetensors",
+    "transformer.blocks.19.norm_attn_norm.attn.Wqkv.input_scale": "model-00015-of-00031.safetensors",
+    "transformer.blocks.19.norm_attn_norm.attn.Wqkv.weight": "model-00015-of-00031.safetensors",
+    "transformer.blocks.19.norm_attn_norm.attn.Wqkv.weight_scale": "model-00015-of-00031.safetensors",
+    "transformer.blocks.19.norm_attn_norm.attn.kv_scale": "model-00015-of-00031.safetensors",
+    "transformer.blocks.19.norm_attn_norm.attn.out_proj.input_scale": "model-00015-of-00031.safetensors",
+    "transformer.blocks.19.norm_attn_norm.attn.out_proj.weight": "model-00015-of-00031.safetensors",
+    "transformer.blocks.19.norm_attn_norm.attn.out_proj.weight_scale": "model-00015-of-00031.safetensors",
+    "transformer.blocks.19.norm_attn_norm.norm_1.weight": "model-00015-of-00031.safetensors",
+    "transformer.blocks.19.norm_attn_norm.norm_2.weight": "model-00015-of-00031.safetensors",
+    "transformer.blocks.2.ffn.experts.mlp.v1.input_scale": "model-00003-of-00031.safetensors",
+    "transformer.blocks.2.ffn.experts.mlp.v1.weight": "model-00003-of-00031.safetensors",
+    "transformer.blocks.2.ffn.experts.mlp.v1.weight_scale": "model-00003-of-00031.safetensors",
+    "transformer.blocks.2.ffn.experts.mlp.w1.input_scale": "model-00002-of-00031.safetensors",
+    "transformer.blocks.2.ffn.experts.mlp.w1.weight": "model-00002-of-00031.safetensors",
+    "transformer.blocks.2.ffn.experts.mlp.w1.weight_scale": "model-00002-of-00031.safetensors",
+    "transformer.blocks.2.ffn.experts.mlp.w2.input_scale": "model-00003-of-00031.safetensors",
+    "transformer.blocks.2.ffn.experts.mlp.w2.weight": "model-00003-of-00031.safetensors",
+    "transformer.blocks.2.ffn.experts.mlp.w2.weight_scale": "model-00003-of-00031.safetensors",
+    "transformer.blocks.2.ffn.router.layer.weight": "model-00002-of-00031.safetensors",
+    "transformer.blocks.2.norm_attn_norm.attn.Wqkv.input_scale": "model-00002-of-00031.safetensors",
+    "transformer.blocks.2.norm_attn_norm.attn.Wqkv.weight": "model-00002-of-00031.safetensors",
+    "transformer.blocks.2.norm_attn_norm.attn.Wqkv.weight_scale": "model-00002-of-00031.safetensors",
+    "transformer.blocks.2.norm_attn_norm.attn.kv_scale": "model-00002-of-00031.safetensors",
+    "transformer.blocks.2.norm_attn_norm.attn.out_proj.input_scale": "model-00002-of-00031.safetensors",
+    "transformer.blocks.2.norm_attn_norm.attn.out_proj.weight": "model-00002-of-00031.safetensors",
+    "transformer.blocks.2.norm_attn_norm.attn.out_proj.weight_scale": "model-00002-of-00031.safetensors",
+    "transformer.blocks.2.norm_attn_norm.norm_1.weight": "model-00002-of-00031.safetensors",
+    "transformer.blocks.2.norm_attn_norm.norm_2.weight": "model-00002-of-00031.safetensors",
+    "transformer.blocks.20.ffn.experts.mlp.v1.input_scale": "model-00016-of-00031.safetensors",
+    "transformer.blocks.20.ffn.experts.mlp.v1.weight": "model-00016-of-00031.safetensors",
+    "transformer.blocks.20.ffn.experts.mlp.v1.weight_scale": "model-00016-of-00031.safetensors",
+    "transformer.blocks.20.ffn.experts.mlp.w1.input_scale": "model-00016-of-00031.safetensors",
+    "transformer.blocks.20.ffn.experts.mlp.w1.weight": "model-00016-of-00031.safetensors",
+    "transformer.blocks.20.ffn.experts.mlp.w1.weight_scale": "model-00016-of-00031.safetensors",
+    "transformer.blocks.20.ffn.experts.mlp.w2.input_scale": "model-00016-of-00031.safetensors",
+    "transformer.blocks.20.ffn.experts.mlp.w2.weight": "model-00016-of-00031.safetensors",
+    "transformer.blocks.20.ffn.experts.mlp.w2.weight_scale": "model-00016-of-00031.safetensors",
+    "transformer.blocks.20.ffn.router.layer.weight": "model-00016-of-00031.safetensors",
+    "transformer.blocks.20.norm_attn_norm.attn.Wqkv.input_scale": "model-00016-of-00031.safetensors",
+    "transformer.blocks.20.norm_attn_norm.attn.Wqkv.weight": "model-00016-of-00031.safetensors",
+    "transformer.blocks.20.norm_attn_norm.attn.Wqkv.weight_scale": "model-00016-of-00031.safetensors",
+    "transformer.blocks.20.norm_attn_norm.attn.kv_scale": "model-00016-of-00031.safetensors",
+    "transformer.blocks.20.norm_attn_norm.attn.out_proj.input_scale": "model-00016-of-00031.safetensors",
+    "transformer.blocks.20.norm_attn_norm.attn.out_proj.weight": "model-00016-of-00031.safetensors",
+    "transformer.blocks.20.norm_attn_norm.attn.out_proj.weight_scale": "model-00016-of-00031.safetensors",
+    "transformer.blocks.20.norm_attn_norm.norm_1.weight": "model-00016-of-00031.safetensors",
+    "transformer.blocks.20.norm_attn_norm.norm_2.weight": "model-00016-of-00031.safetensors",
+    "transformer.blocks.21.ffn.experts.mlp.v1.input_scale": "model-00017-of-00031.safetensors",
+    "transformer.blocks.21.ffn.experts.mlp.v1.weight": "model-00017-of-00031.safetensors",
+    "transformer.blocks.21.ffn.experts.mlp.v1.weight_scale": "model-00017-of-00031.safetensors",
+    "transformer.blocks.21.ffn.experts.mlp.w1.input_scale": "model-00017-of-00031.safetensors",
+    "transformer.blocks.21.ffn.experts.mlp.w1.weight": "model-00017-of-00031.safetensors",
+    "transformer.blocks.21.ffn.experts.mlp.w1.weight_scale": "model-00017-of-00031.safetensors",
+    "transformer.blocks.21.ffn.experts.mlp.w2.input_scale": "model-00017-of-00031.safetensors",
+    "transformer.blocks.21.ffn.experts.mlp.w2.weight": "model-00017-of-00031.safetensors",
+    "transformer.blocks.21.ffn.experts.mlp.w2.weight_scale": "model-00017-of-00031.safetensors",
+    "transformer.blocks.21.ffn.router.layer.weight": "model-00016-of-00031.safetensors",
+    "transformer.blocks.21.norm_attn_norm.attn.Wqkv.input_scale": "model-00016-of-00031.safetensors",
+    "transformer.blocks.21.norm_attn_norm.attn.Wqkv.weight": "model-00016-of-00031.safetensors",
+    "transformer.blocks.21.norm_attn_norm.attn.Wqkv.weight_scale": "model-00016-of-00031.safetensors",
+    "transformer.blocks.21.norm_attn_norm.attn.kv_scale": "model-00016-of-00031.safetensors",
+    "transformer.blocks.21.norm_attn_norm.attn.out_proj.input_scale": "model-00016-of-00031.safetensors",
+    "transformer.blocks.21.norm_attn_norm.attn.out_proj.weight": "model-00016-of-00031.safetensors",
+    "transformer.blocks.21.norm_attn_norm.attn.out_proj.weight_scale": "model-00016-of-00031.safetensors",
+    "transformer.blocks.21.norm_attn_norm.norm_1.weight": "model-00016-of-00031.safetensors",
+    "transformer.blocks.21.norm_attn_norm.norm_2.weight": "model-00016-of-00031.safetensors",
+    "transformer.blocks.22.ffn.experts.mlp.v1.input_scale": "model-00018-of-00031.safetensors",
+    "transformer.blocks.22.ffn.experts.mlp.v1.weight": "model-00018-of-00031.safetensors",
+    "transformer.blocks.22.ffn.experts.mlp.v1.weight_scale": "model-00018-of-00031.safetensors",
+    "transformer.blocks.22.ffn.experts.mlp.w1.input_scale": "model-00017-of-00031.safetensors",
+    "transformer.blocks.22.ffn.experts.mlp.w1.weight": "model-00017-of-00031.safetensors",
+    "transformer.blocks.22.ffn.experts.mlp.w1.weight_scale": "model-00017-of-00031.safetensors",
+    "transformer.blocks.22.ffn.experts.mlp.w2.input_scale": "model-00018-of-00031.safetensors",
+    "transformer.blocks.22.ffn.experts.mlp.w2.weight": "model-00018-of-00031.safetensors",
+    "transformer.blocks.22.ffn.experts.mlp.w2.weight_scale": "model-00018-of-00031.safetensors",
+    "transformer.blocks.22.ffn.router.layer.weight": "model-00017-of-00031.safetensors",
+    "transformer.blocks.22.norm_attn_norm.attn.Wqkv.input_scale": "model-00017-of-00031.safetensors",
+    "transformer.blocks.22.norm_attn_norm.attn.Wqkv.weight": "model-00017-of-00031.safetensors",
+    "transformer.blocks.22.norm_attn_norm.attn.Wqkv.weight_scale": "model-00017-of-00031.safetensors",
+    "transformer.blocks.22.norm_attn_norm.attn.kv_scale": "model-00017-of-00031.safetensors",
+    "transformer.blocks.22.norm_attn_norm.attn.out_proj.input_scale": "model-00017-of-00031.safetensors",
+    "transformer.blocks.22.norm_attn_norm.attn.out_proj.weight": "model-00017-of-00031.safetensors",
+    "transformer.blocks.22.norm_attn_norm.attn.out_proj.weight_scale": "model-00017-of-00031.safetensors",
+    "transformer.blocks.22.norm_attn_norm.norm_1.weight": "model-00017-of-00031.safetensors",
+    "transformer.blocks.22.norm_attn_norm.norm_2.weight": "model-00017-of-00031.safetensors",
+    "transformer.blocks.23.ffn.experts.mlp.v1.input_scale": "model-00018-of-00031.safetensors",
+    "transformer.blocks.23.ffn.experts.mlp.v1.weight": "model-00018-of-00031.safetensors",
+    "transformer.blocks.23.ffn.experts.mlp.v1.weight_scale": "model-00018-of-00031.safetensors",
+    "transformer.blocks.23.ffn.experts.mlp.w1.input_scale": "model-00018-of-00031.safetensors",
+    "transformer.blocks.23.ffn.experts.mlp.w1.weight": "model-00018-of-00031.safetensors",
+    "transformer.blocks.23.ffn.experts.mlp.w1.weight_scale": "model-00018-of-00031.safetensors",
+    "transformer.blocks.23.ffn.experts.mlp.w2.input_scale": "model-00019-of-00031.safetensors",
+    "transformer.blocks.23.ffn.experts.mlp.w2.weight": "model-00019-of-00031.safetensors",
+    "transformer.blocks.23.ffn.experts.mlp.w2.weight_scale": "model-00019-of-00031.safetensors",
+    "transformer.blocks.23.ffn.router.layer.weight": "model-00018-of-00031.safetensors",
+    "transformer.blocks.23.norm_attn_norm.attn.Wqkv.input_scale": "model-00018-of-00031.safetensors",
+    "transformer.blocks.23.norm_attn_norm.attn.Wqkv.weight": "model-00018-of-00031.safetensors",
+    "transformer.blocks.23.norm_attn_norm.attn.Wqkv.weight_scale": "model-00018-of-00031.safetensors",
+    "transformer.blocks.23.norm_attn_norm.attn.kv_scale": "model-00018-of-00031.safetensors",
+    "transformer.blocks.23.norm_attn_norm.attn.out_proj.input_scale": "model-00018-of-00031.safetensors",
+    "transformer.blocks.23.norm_attn_norm.attn.out_proj.weight": "model-00018-of-00031.safetensors",
+    "transformer.blocks.23.norm_attn_norm.attn.out_proj.weight_scale": "model-00018-of-00031.safetensors",
+    "transformer.blocks.23.norm_attn_norm.norm_1.weight": "model-00018-of-00031.safetensors",
+    "transformer.blocks.23.norm_attn_norm.norm_2.weight": "model-00018-of-00031.safetensors",
+    "transformer.blocks.24.ffn.experts.mlp.v1.input_scale": "model-00019-of-00031.safetensors",
+    "transformer.blocks.24.ffn.experts.mlp.v1.weight": "model-00019-of-00031.safetensors",
+    "transformer.blocks.24.ffn.experts.mlp.v1.weight_scale": "model-00019-of-00031.safetensors",
+    "transformer.blocks.24.ffn.experts.mlp.w1.input_scale": "model-00019-of-00031.safetensors",
+    "transformer.blocks.24.ffn.experts.mlp.w1.weight": "model-00019-of-00031.safetensors",
+    "transformer.blocks.24.ffn.experts.mlp.w1.weight_scale": "model-00019-of-00031.safetensors",
+    "transformer.blocks.24.ffn.experts.mlp.w2.input_scale": "model-00019-of-00031.safetensors",
+    "transformer.blocks.24.ffn.experts.mlp.w2.weight": "model-00019-of-00031.safetensors",
+    "transformer.blocks.24.ffn.experts.mlp.w2.weight_scale": "model-00019-of-00031.safetensors",
+    "transformer.blocks.24.ffn.router.layer.weight": "model-00019-of-00031.safetensors",
+    "transformer.blocks.24.norm_attn_norm.attn.Wqkv.input_scale": "model-00019-of-00031.safetensors",
+    "transformer.blocks.24.norm_attn_norm.attn.Wqkv.weight": "model-00019-of-00031.safetensors",
+    "transformer.blocks.24.norm_attn_norm.attn.Wqkv.weight_scale": "model-00019-of-00031.safetensors",
+    "transformer.blocks.24.norm_attn_norm.attn.kv_scale": "model-00019-of-00031.safetensors",
+    "transformer.blocks.24.norm_attn_norm.attn.out_proj.input_scale": "model-00019-of-00031.safetensors",
+    "transformer.blocks.24.norm_attn_norm.attn.out_proj.weight": "model-00019-of-00031.safetensors",
+    "transformer.blocks.24.norm_attn_norm.attn.out_proj.weight_scale": "model-00019-of-00031.safetensors",
+    "transformer.blocks.24.norm_attn_norm.norm_1.weight": "model-00019-of-00031.safetensors",
+    "transformer.blocks.24.norm_attn_norm.norm_2.weight": "model-00019-of-00031.safetensors",
+    "transformer.blocks.25.ffn.experts.mlp.v1.input_scale": "model-00020-of-00031.safetensors",
+    "transformer.blocks.25.ffn.experts.mlp.v1.weight": "model-00020-of-00031.safetensors",
+    "transformer.blocks.25.ffn.experts.mlp.v1.weight_scale": "model-00020-of-00031.safetensors",
+    "transformer.blocks.25.ffn.experts.mlp.w1.input_scale": "model-00020-of-00031.safetensors",
+    "transformer.blocks.25.ffn.experts.mlp.w1.weight": "model-00020-of-00031.safetensors",
+    "transformer.blocks.25.ffn.experts.mlp.w1.weight_scale": "model-00020-of-00031.safetensors",
+    "transformer.blocks.25.ffn.experts.mlp.w2.input_scale": "model-00020-of-00031.safetensors",
+    "transformer.blocks.25.ffn.experts.mlp.w2.weight": "model-00020-of-00031.safetensors",
+    "transformer.blocks.25.ffn.experts.mlp.w2.weight_scale": "model-00020-of-00031.safetensors",
+    "transformer.blocks.25.ffn.router.layer.weight": "model-00019-of-00031.safetensors",
+    "transformer.blocks.25.norm_attn_norm.attn.Wqkv.input_scale": "model-00019-of-00031.safetensors",
+    "transformer.blocks.25.norm_attn_norm.attn.Wqkv.weight": "model-00019-of-00031.safetensors",
+    "transformer.blocks.25.norm_attn_norm.attn.Wqkv.weight_scale": "model-00019-of-00031.safetensors",
+    "transformer.blocks.25.norm_attn_norm.attn.kv_scale": "model-00019-of-00031.safetensors",
+    "transformer.blocks.25.norm_attn_norm.attn.out_proj.input_scale": "model-00019-of-00031.safetensors",
+    "transformer.blocks.25.norm_attn_norm.attn.out_proj.weight": "model-00019-of-00031.safetensors",
+    "transformer.blocks.25.norm_attn_norm.attn.out_proj.weight_scale": "model-00019-of-00031.safetensors",
+    "transformer.blocks.25.norm_attn_norm.norm_1.weight": "model-00019-of-00031.safetensors",
+    "transformer.blocks.25.norm_attn_norm.norm_2.weight": "model-00019-of-00031.safetensors",
+    "transformer.blocks.26.ffn.experts.mlp.v1.input_scale": "model-00021-of-00031.safetensors",
+    "transformer.blocks.26.ffn.experts.mlp.v1.weight": "model-00021-of-00031.safetensors",
+    "transformer.blocks.26.ffn.experts.mlp.v1.weight_scale": "model-00021-of-00031.safetensors",
+    "transformer.blocks.26.ffn.experts.mlp.w1.input_scale": "model-00020-of-00031.safetensors",
+    "transformer.blocks.26.ffn.experts.mlp.w1.weight": "model-00020-of-00031.safetensors",
+    "transformer.blocks.26.ffn.experts.mlp.w1.weight_scale": "model-00020-of-00031.safetensors",
+    "transformer.blocks.26.ffn.experts.mlp.w2.input_scale": "model-00021-of-00031.safetensors",
+    "transformer.blocks.26.ffn.experts.mlp.w2.weight": "model-00021-of-00031.safetensors",
+    "transformer.blocks.26.ffn.experts.mlp.w2.weight_scale": "model-00021-of-00031.safetensors",
+    "transformer.blocks.26.ffn.router.layer.weight": "model-00020-of-00031.safetensors",
+    "transformer.blocks.26.norm_attn_norm.attn.Wqkv.input_scale": "model-00020-of-00031.safetensors",
+    "transformer.blocks.26.norm_attn_norm.attn.Wqkv.weight": "model-00020-of-00031.safetensors",
+    "transformer.blocks.26.norm_attn_norm.attn.Wqkv.weight_scale": "model-00020-of-00031.safetensors",
+    "transformer.blocks.26.norm_attn_norm.attn.kv_scale": "model-00020-of-00031.safetensors",
+    "transformer.blocks.26.norm_attn_norm.attn.out_proj.input_scale": "model-00020-of-00031.safetensors",
+    "transformer.blocks.26.norm_attn_norm.attn.out_proj.weight": "model-00020-of-00031.safetensors",
+    "transformer.blocks.26.norm_attn_norm.attn.out_proj.weight_scale": "model-00020-of-00031.safetensors",
+    "transformer.blocks.26.norm_attn_norm.norm_1.weight": "model-00020-of-00031.safetensors",
+    "transformer.blocks.26.norm_attn_norm.norm_2.weight": "model-00020-of-00031.safetensors",
+    "transformer.blocks.27.ffn.experts.mlp.v1.input_scale": "model-00021-of-00031.safetensors",
+    "transformer.blocks.27.ffn.experts.mlp.v1.weight": "model-00021-of-00031.safetensors",
+    "transformer.blocks.27.ffn.experts.mlp.v1.weight_scale": "model-00021-of-00031.safetensors",
+    "transformer.blocks.27.ffn.experts.mlp.w1.input_scale": "model-00021-of-00031.safetensors",
+    "transformer.blocks.27.ffn.experts.mlp.w1.weight": "model-00021-of-00031.safetensors",
+    "transformer.blocks.27.ffn.experts.mlp.w1.weight_scale": "model-00021-of-00031.safetensors",
+    "transformer.blocks.27.ffn.experts.mlp.w2.input_scale": "model-00022-of-00031.safetensors",
+    "transformer.blocks.27.ffn.experts.mlp.w2.weight": "model-00022-of-00031.safetensors",
+    "transformer.blocks.27.ffn.experts.mlp.w2.weight_scale": "model-00022-of-00031.safetensors",
+    "transformer.blocks.27.ffn.router.layer.weight": "model-00021-of-00031.safetensors",
+    "transformer.blocks.27.norm_attn_norm.attn.Wqkv.input_scale": "model-00021-of-00031.safetensors",
+    "transformer.blocks.27.norm_attn_norm.attn.Wqkv.weight": "model-00021-of-00031.safetensors",
+    "transformer.blocks.27.norm_attn_norm.attn.Wqkv.weight_scale": "model-00021-of-00031.safetensors",
+    "transformer.blocks.27.norm_attn_norm.attn.kv_scale": "model-00021-of-00031.safetensors",
+    "transformer.blocks.27.norm_attn_norm.attn.out_proj.input_scale": "model-00021-of-00031.safetensors",
+    "transformer.blocks.27.norm_attn_norm.attn.out_proj.weight": "model-00021-of-00031.safetensors",
+    "transformer.blocks.27.norm_attn_norm.attn.out_proj.weight_scale": "model-00021-of-00031.safetensors",
+    "transformer.blocks.27.norm_attn_norm.norm_1.weight": "model-00021-of-00031.safetensors",
+    "transformer.blocks.27.norm_attn_norm.norm_2.weight": "model-00021-of-00031.safetensors",
+    "transformer.blocks.28.ffn.experts.mlp.v1.input_scale": "model-00022-of-00031.safetensors",
+    "transformer.blocks.28.ffn.experts.mlp.v1.weight": "model-00022-of-00031.safetensors",
+    "transformer.blocks.28.ffn.experts.mlp.v1.weight_scale": "model-00022-of-00031.safetensors",
+    "transformer.blocks.28.ffn.experts.mlp.w1.input_scale": "model-00022-of-00031.safetensors",
+    "transformer.blocks.28.ffn.experts.mlp.w1.weight": "model-00022-of-00031.safetensors",
+    "transformer.blocks.28.ffn.experts.mlp.w1.weight_scale": "model-00022-of-00031.safetensors",
+    "transformer.blocks.28.ffn.experts.mlp.w2.input_scale": "model-00022-of-00031.safetensors",
+    "transformer.blocks.28.ffn.experts.mlp.w2.weight": "model-00022-of-00031.safetensors",
+    "transformer.blocks.28.ffn.experts.mlp.w2.weight_scale": "model-00022-of-00031.safetensors",
+    "transformer.blocks.28.ffn.router.layer.weight": "model-00022-of-00031.safetensors",
+    "transformer.blocks.28.norm_attn_norm.attn.Wqkv.input_scale": "model-00022-of-00031.safetensors",
+    "transformer.blocks.28.norm_attn_norm.attn.Wqkv.weight": "model-00022-of-00031.safetensors",
+    "transformer.blocks.28.norm_attn_norm.attn.Wqkv.weight_scale": "model-00022-of-00031.safetensors",
+    "transformer.blocks.28.norm_attn_norm.attn.kv_scale": "model-00022-of-00031.safetensors",
+    "transformer.blocks.28.norm_attn_norm.attn.out_proj.input_scale": "model-00022-of-00031.safetensors",
+    "transformer.blocks.28.norm_attn_norm.attn.out_proj.weight": "model-00022-of-00031.safetensors",
+    "transformer.blocks.28.norm_attn_norm.attn.out_proj.weight_scale": "model-00022-of-00031.safetensors",
+    "transformer.blocks.28.norm_attn_norm.norm_1.weight": "model-00022-of-00031.safetensors",
+    "transformer.blocks.28.norm_attn_norm.norm_2.weight": "model-00022-of-00031.safetensors",
+    "transformer.blocks.29.ffn.experts.mlp.v1.input_scale": "model-00023-of-00031.safetensors",
+    "transformer.blocks.29.ffn.experts.mlp.v1.weight": "model-00023-of-00031.safetensors",
+    "transformer.blocks.29.ffn.experts.mlp.v1.weight_scale": "model-00023-of-00031.safetensors",
+    "transformer.blocks.29.ffn.experts.mlp.w1.input_scale": "model-00023-of-00031.safetensors",
+    "transformer.blocks.29.ffn.experts.mlp.w1.weight": "model-00023-of-00031.safetensors",
+    "transformer.blocks.29.ffn.experts.mlp.w1.weight_scale": "model-00023-of-00031.safetensors",
+    "transformer.blocks.29.ffn.experts.mlp.w2.input_scale": "model-00023-of-00031.safetensors",
+    "transformer.blocks.29.ffn.experts.mlp.w2.weight": "model-00023-of-00031.safetensors",
+    "transformer.blocks.29.ffn.experts.mlp.w2.weight_scale": "model-00023-of-00031.safetensors",
+    "transformer.blocks.29.ffn.router.layer.weight": "model-00022-of-00031.safetensors",
+    "transformer.blocks.29.norm_attn_norm.attn.Wqkv.input_scale": "model-00022-of-00031.safetensors",
+    "transformer.blocks.29.norm_attn_norm.attn.Wqkv.weight": "model-00022-of-00031.safetensors",
+    "transformer.blocks.29.norm_attn_norm.attn.Wqkv.weight_scale": "model-00022-of-00031.safetensors",
+    "transformer.blocks.29.norm_attn_norm.attn.kv_scale": "model-00022-of-00031.safetensors",
+    "transformer.blocks.29.norm_attn_norm.attn.out_proj.input_scale": "model-00022-of-00031.safetensors",
+    "transformer.blocks.29.norm_attn_norm.attn.out_proj.weight": "model-00022-of-00031.safetensors",
+    "transformer.blocks.29.norm_attn_norm.attn.out_proj.weight_scale": "model-00022-of-00031.safetensors",
+    "transformer.blocks.29.norm_attn_norm.norm_1.weight": "model-00022-of-00031.safetensors",
+    "transformer.blocks.29.norm_attn_norm.norm_2.weight": "model-00022-of-00031.safetensors",
+    "transformer.blocks.3.ffn.experts.mlp.v1.input_scale": "model-00003-of-00031.safetensors",
+    "transformer.blocks.3.ffn.experts.mlp.v1.weight": "model-00003-of-00031.safetensors",
+    "transformer.blocks.3.ffn.experts.mlp.v1.weight_scale": "model-00003-of-00031.safetensors",
+    "transformer.blocks.3.ffn.experts.mlp.w1.input_scale": "model-00003-of-00031.safetensors",
+    "transformer.blocks.3.ffn.experts.mlp.w1.weight": "model-00003-of-00031.safetensors",
+    "transformer.blocks.3.ffn.experts.mlp.w1.weight_scale": "model-00003-of-00031.safetensors",
+    "transformer.blocks.3.ffn.experts.mlp.w2.input_scale": "model-00004-of-00031.safetensors",
+    "transformer.blocks.3.ffn.experts.mlp.w2.weight": "model-00004-of-00031.safetensors",
+    "transformer.blocks.3.ffn.experts.mlp.w2.weight_scale": "model-00004-of-00031.safetensors",
+    "transformer.blocks.3.ffn.router.layer.weight": "model-00003-of-00031.safetensors",
+    "transformer.blocks.3.norm_attn_norm.attn.Wqkv.input_scale": "model-00003-of-00031.safetensors",
+    "transformer.blocks.3.norm_attn_norm.attn.Wqkv.weight": "model-00003-of-00031.safetensors",
+    "transformer.blocks.3.norm_attn_norm.attn.Wqkv.weight_scale": "model-00003-of-00031.safetensors",
+    "transformer.blocks.3.norm_attn_norm.attn.kv_scale": "model-00003-of-00031.safetensors",
+    "transformer.blocks.3.norm_attn_norm.attn.out_proj.input_scale": "model-00003-of-00031.safetensors",
+    "transformer.blocks.3.norm_attn_norm.attn.out_proj.weight": "model-00003-of-00031.safetensors",
+    "transformer.blocks.3.norm_attn_norm.attn.out_proj.weight_scale": "model-00003-of-00031.safetensors",
+    "transformer.blocks.3.norm_attn_norm.norm_1.weight": "model-00003-of-00031.safetensors",
+    "transformer.blocks.3.norm_attn_norm.norm_2.weight": "model-00003-of-00031.safetensors",
+    "transformer.blocks.30.ffn.experts.mlp.v1.input_scale": "model-00024-of-00031.safetensors",
+    "transformer.blocks.30.ffn.experts.mlp.v1.weight": "model-00024-of-00031.safetensors",
+    "transformer.blocks.30.ffn.experts.mlp.v1.weight_scale": "model-00024-of-00031.safetensors",
+    "transformer.blocks.30.ffn.experts.mlp.w1.input_scale": "model-00023-of-00031.safetensors",
+    "transformer.blocks.30.ffn.experts.mlp.w1.weight": "model-00023-of-00031.safetensors",
+    "transformer.blocks.30.ffn.experts.mlp.w1.weight_scale": "model-00023-of-00031.safetensors",
+    "transformer.blocks.30.ffn.experts.mlp.w2.input_scale": "model-00024-of-00031.safetensors",
+    "transformer.blocks.30.ffn.experts.mlp.w2.weight": "model-00024-of-00031.safetensors",
+    "transformer.blocks.30.ffn.experts.mlp.w2.weight_scale": "model-00024-of-00031.safetensors",
+    "transformer.blocks.30.ffn.router.layer.weight": "model-00023-of-00031.safetensors",
+    "transformer.blocks.30.norm_attn_norm.attn.Wqkv.input_scale": "model-00023-of-00031.safetensors",
+    "transformer.blocks.30.norm_attn_norm.attn.Wqkv.weight": "model-00023-of-00031.safetensors",
+    "transformer.blocks.30.norm_attn_norm.attn.Wqkv.weight_scale": "model-00023-of-00031.safetensors",
+    "transformer.blocks.30.norm_attn_norm.attn.kv_scale": "model-00023-of-00031.safetensors",
+    "transformer.blocks.30.norm_attn_norm.attn.out_proj.input_scale": "model-00023-of-00031.safetensors",
+    "transformer.blocks.30.norm_attn_norm.attn.out_proj.weight": "model-00023-of-00031.safetensors",
+    "transformer.blocks.30.norm_attn_norm.attn.out_proj.weight_scale": "model-00023-of-00031.safetensors",
+    "transformer.blocks.30.norm_attn_norm.norm_1.weight": "model-00023-of-00031.safetensors",
+    "transformer.blocks.30.norm_attn_norm.norm_2.weight": "model-00023-of-00031.safetensors",
+    "transformer.blocks.31.ffn.experts.mlp.v1.input_scale": "model-00024-of-00031.safetensors",
+    "transformer.blocks.31.ffn.experts.mlp.v1.weight": "model-00024-of-00031.safetensors",
+    "transformer.blocks.31.ffn.experts.mlp.v1.weight_scale": "model-00024-of-00031.safetensors",
+    "transformer.blocks.31.ffn.experts.mlp.w1.input_scale": "model-00024-of-00031.safetensors",
+    "transformer.blocks.31.ffn.experts.mlp.w1.weight": "model-00024-of-00031.safetensors",
+    "transformer.blocks.31.ffn.experts.mlp.w1.weight_scale": "model-00024-of-00031.safetensors",
+    "transformer.blocks.31.ffn.experts.mlp.w2.input_scale": "model-00025-of-00031.safetensors",
+    "transformer.blocks.31.ffn.experts.mlp.w2.weight": "model-00025-of-00031.safetensors",
+    "transformer.blocks.31.ffn.experts.mlp.w2.weight_scale": "model-00025-of-00031.safetensors",
+    "transformer.blocks.31.ffn.router.layer.weight": "model-00024-of-00031.safetensors",
+    "transformer.blocks.31.norm_attn_norm.attn.Wqkv.input_scale": "model-00024-of-00031.safetensors",
+    "transformer.blocks.31.norm_attn_norm.attn.Wqkv.weight": "model-00024-of-00031.safetensors",
+    "transformer.blocks.31.norm_attn_norm.attn.Wqkv.weight_scale": "model-00024-of-00031.safetensors",
+    "transformer.blocks.31.norm_attn_norm.attn.kv_scale": "model-00024-of-00031.safetensors",
+    "transformer.blocks.31.norm_attn_norm.attn.out_proj.input_scale": "model-00024-of-00031.safetensors",
+    "transformer.blocks.31.norm_attn_norm.attn.out_proj.weight": "model-00024-of-00031.safetensors",
+    "transformer.blocks.31.norm_attn_norm.attn.out_proj.weight_scale": "model-00024-of-00031.safetensors",
+    "transformer.blocks.31.norm_attn_norm.norm_1.weight": "model-00024-of-00031.safetensors",
+    "transformer.blocks.31.norm_attn_norm.norm_2.weight": "model-00024-of-00031.safetensors",
+    "transformer.blocks.32.ffn.experts.mlp.v1.input_scale": "model-00025-of-00031.safetensors",
+    "transformer.blocks.32.ffn.experts.mlp.v1.weight": "model-00025-of-00031.safetensors",
+    "transformer.blocks.32.ffn.experts.mlp.v1.weight_scale": "model-00025-of-00031.safetensors",
+    "transformer.blocks.32.ffn.experts.mlp.w1.input_scale": "model-00025-of-00031.safetensors",
+    "transformer.blocks.32.ffn.experts.mlp.w1.weight": "model-00025-of-00031.safetensors",
+    "transformer.blocks.32.ffn.experts.mlp.w1.weight_scale": "model-00025-of-00031.safetensors",
+    "transformer.blocks.32.ffn.experts.mlp.w2.input_scale": "model-00025-of-00031.safetensors",
+    "transformer.blocks.32.ffn.experts.mlp.w2.weight": "model-00025-of-00031.safetensors",
+    "transformer.blocks.32.ffn.experts.mlp.w2.weight_scale": "model-00025-of-00031.safetensors",
+    "transformer.blocks.32.ffn.router.layer.weight": "model-00025-of-00031.safetensors",
+    "transformer.blocks.32.norm_attn_norm.attn.Wqkv.input_scale": "model-00025-of-00031.safetensors",
+    "transformer.blocks.32.norm_attn_norm.attn.Wqkv.weight": "model-00025-of-00031.safetensors",
+    "transformer.blocks.32.norm_attn_norm.attn.Wqkv.weight_scale": "model-00025-of-00031.safetensors",
+    "transformer.blocks.32.norm_attn_norm.attn.kv_scale": "model-00025-of-00031.safetensors",
+    "transformer.blocks.32.norm_attn_norm.attn.out_proj.input_scale": "model-00025-of-00031.safetensors",
+    "transformer.blocks.32.norm_attn_norm.attn.out_proj.weight": "model-00025-of-00031.safetensors",
+    "transformer.blocks.32.norm_attn_norm.attn.out_proj.weight_scale": "model-00025-of-00031.safetensors",
+    "transformer.blocks.32.norm_attn_norm.norm_1.weight": "model-00025-of-00031.safetensors",
+    "transformer.blocks.32.norm_attn_norm.norm_2.weight": "model-00025-of-00031.safetensors",
+    "transformer.blocks.33.ffn.experts.mlp.v1.input_scale": "model-00026-of-00031.safetensors",
+    "transformer.blocks.33.ffn.experts.mlp.v1.weight": "model-00026-of-00031.safetensors",
+    "transformer.blocks.33.ffn.experts.mlp.v1.weight_scale": "model-00026-of-00031.safetensors",
+    "transformer.blocks.33.ffn.experts.mlp.w1.input_scale": "model-00026-of-00031.safetensors",
+    "transformer.blocks.33.ffn.experts.mlp.w1.weight": "model-00026-of-00031.safetensors",
+    "transformer.blocks.33.ffn.experts.mlp.w1.weight_scale": "model-00026-of-00031.safetensors",
+    "transformer.blocks.33.ffn.experts.mlp.w2.input_scale": "model-00026-of-00031.safetensors",
+    "transformer.blocks.33.ffn.experts.mlp.w2.weight": "model-00026-of-00031.safetensors",
+    "transformer.blocks.33.ffn.experts.mlp.w2.weight_scale": "model-00026-of-00031.safetensors",
+    "transformer.blocks.33.ffn.router.layer.weight": "model-00025-of-00031.safetensors",
+    "transformer.blocks.33.norm_attn_norm.attn.Wqkv.input_scale": "model-00025-of-00031.safetensors",
+    "transformer.blocks.33.norm_attn_norm.attn.Wqkv.weight": "model-00025-of-00031.safetensors",
+    "transformer.blocks.33.norm_attn_norm.attn.Wqkv.weight_scale": "model-00025-of-00031.safetensors",
+    "transformer.blocks.33.norm_attn_norm.attn.kv_scale": "model-00025-of-00031.safetensors",
+    "transformer.blocks.33.norm_attn_norm.attn.out_proj.input_scale": "model-00025-of-00031.safetensors",
+    "transformer.blocks.33.norm_attn_norm.attn.out_proj.weight": "model-00025-of-00031.safetensors",
+    "transformer.blocks.33.norm_attn_norm.attn.out_proj.weight_scale": "model-00025-of-00031.safetensors",
+    "transformer.blocks.33.norm_attn_norm.norm_1.weight": "model-00025-of-00031.safetensors",
+    "transformer.blocks.33.norm_attn_norm.norm_2.weight": "model-00025-of-00031.safetensors",
+    "transformer.blocks.34.ffn.experts.mlp.v1.input_scale": "model-00027-of-00031.safetensors",
+    "transformer.blocks.34.ffn.experts.mlp.v1.weight": "model-00027-of-00031.safetensors",
+    "transformer.blocks.34.ffn.experts.mlp.v1.weight_scale": "model-00027-of-00031.safetensors",
+    "transformer.blocks.34.ffn.experts.mlp.w1.input_scale": "model-00026-of-00031.safetensors",
+    "transformer.blocks.34.ffn.experts.mlp.w1.weight": "model-00026-of-00031.safetensors",
+    "transformer.blocks.34.ffn.experts.mlp.w1.weight_scale": "model-00026-of-00031.safetensors",
+    "transformer.blocks.34.ffn.experts.mlp.w2.input_scale": "model-00027-of-00031.safetensors",
+    "transformer.blocks.34.ffn.experts.mlp.w2.weight": "model-00027-of-00031.safetensors",
+    "transformer.blocks.34.ffn.experts.mlp.w2.weight_scale": "model-00027-of-00031.safetensors",
+    "transformer.blocks.34.ffn.router.layer.weight": "model-00026-of-00031.safetensors",
+    "transformer.blocks.34.norm_attn_norm.attn.Wqkv.input_scale": "model-00026-of-00031.safetensors",
+    "transformer.blocks.34.norm_attn_norm.attn.Wqkv.weight": "model-00026-of-00031.safetensors",
+    "transformer.blocks.34.norm_attn_norm.attn.Wqkv.weight_scale": "model-00026-of-00031.safetensors",
+    "transformer.blocks.34.norm_attn_norm.attn.kv_scale": "model-00026-of-00031.safetensors",
+    "transformer.blocks.34.norm_attn_norm.attn.out_proj.input_scale": "model-00026-of-00031.safetensors",
+    "transformer.blocks.34.norm_attn_norm.attn.out_proj.weight": "model-00026-of-00031.safetensors",
+    "transformer.blocks.34.norm_attn_norm.attn.out_proj.weight_scale": "model-00026-of-00031.safetensors",
+    "transformer.blocks.34.norm_attn_norm.norm_1.weight": "model-00026-of-00031.safetensors",
+    "transformer.blocks.34.norm_attn_norm.norm_2.weight": "model-00026-of-00031.safetensors",
+    "transformer.blocks.35.ffn.experts.mlp.v1.input_scale": "model-00027-of-00031.safetensors",
+    "transformer.blocks.35.ffn.experts.mlp.v1.weight": "model-00027-of-00031.safetensors",
+    "transformer.blocks.35.ffn.experts.mlp.v1.weight_scale": "model-00027-of-00031.safetensors",
+    "transformer.blocks.35.ffn.experts.mlp.w1.input_scale": "model-00027-of-00031.safetensors",
+    "transformer.blocks.35.ffn.experts.mlp.w1.weight": "model-00027-of-00031.safetensors",
+    "transformer.blocks.35.ffn.experts.mlp.w1.weight_scale": "model-00027-of-00031.safetensors",
+    "transformer.blocks.35.ffn.experts.mlp.w2.input_scale": "model-00028-of-00031.safetensors",
+    "transformer.blocks.35.ffn.experts.mlp.w2.weight": "model-00028-of-00031.safetensors",
+    "transformer.blocks.35.ffn.experts.mlp.w2.weight_scale": "model-00028-of-00031.safetensors",
+    "transformer.blocks.35.ffn.router.layer.weight": "model-00027-of-00031.safetensors",
+    "transformer.blocks.35.norm_attn_norm.attn.Wqkv.input_scale": "model-00027-of-00031.safetensors",
+    "transformer.blocks.35.norm_attn_norm.attn.Wqkv.weight": "model-00027-of-00031.safetensors",
+    "transformer.blocks.35.norm_attn_norm.attn.Wqkv.weight_scale": "model-00027-of-00031.safetensors",
+    "transformer.blocks.35.norm_attn_norm.attn.kv_scale": "model-00027-of-00031.safetensors",
+    "transformer.blocks.35.norm_attn_norm.attn.out_proj.input_scale": "model-00027-of-00031.safetensors",
+    "transformer.blocks.35.norm_attn_norm.attn.out_proj.weight": "model-00027-of-00031.safetensors",
+    "transformer.blocks.35.norm_attn_norm.attn.out_proj.weight_scale": "model-00027-of-00031.safetensors",
+    "transformer.blocks.35.norm_attn_norm.norm_1.weight": "model-00027-of-00031.safetensors",
+    "transformer.blocks.35.norm_attn_norm.norm_2.weight": "model-00027-of-00031.safetensors",
+    "transformer.blocks.36.ffn.experts.mlp.v1.input_scale": "model-00028-of-00031.safetensors",
+    "transformer.blocks.36.ffn.experts.mlp.v1.weight": "model-00028-of-00031.safetensors",
+    "transformer.blocks.36.ffn.experts.mlp.v1.weight_scale": "model-00028-of-00031.safetensors",
+    "transformer.blocks.36.ffn.experts.mlp.w1.input_scale": "model-00028-of-00031.safetensors",
+    "transformer.blocks.36.ffn.experts.mlp.w1.weight": "model-00028-of-00031.safetensors",
+    "transformer.blocks.36.ffn.experts.mlp.w1.weight_scale": "model-00028-of-00031.safetensors",
+    "transformer.blocks.36.ffn.experts.mlp.w2.input_scale": "model-00028-of-00031.safetensors",
+    "transformer.blocks.36.ffn.experts.mlp.w2.weight": "model-00028-of-00031.safetensors",
+    "transformer.blocks.36.ffn.experts.mlp.w2.weight_scale": "model-00028-of-00031.safetensors",
+    "transformer.blocks.36.ffn.router.layer.weight": "model-00028-of-00031.safetensors",
+    "transformer.blocks.36.norm_attn_norm.attn.Wqkv.input_scale": "model-00028-of-00031.safetensors",
+    "transformer.blocks.36.norm_attn_norm.attn.Wqkv.weight": "model-00028-of-00031.safetensors",
+    "transformer.blocks.36.norm_attn_norm.attn.Wqkv.weight_scale": "model-00028-of-00031.safetensors",
+    "transformer.blocks.36.norm_attn_norm.attn.kv_scale": "model-00028-of-00031.safetensors",
+    "transformer.blocks.36.norm_attn_norm.attn.out_proj.input_scale": "model-00028-of-00031.safetensors",
+    "transformer.blocks.36.norm_attn_norm.attn.out_proj.weight": "model-00028-of-00031.safetensors",
+    "transformer.blocks.36.norm_attn_norm.attn.out_proj.weight_scale": "model-00028-of-00031.safetensors",
+    "transformer.blocks.36.norm_attn_norm.norm_1.weight": "model-00028-of-00031.safetensors",
+    "transformer.blocks.36.norm_attn_norm.norm_2.weight": "model-00028-of-00031.safetensors",
+    "transformer.blocks.37.ffn.experts.mlp.v1.input_scale": "model-00029-of-00031.safetensors",
+    "transformer.blocks.37.ffn.experts.mlp.v1.weight": "model-00029-of-00031.safetensors",
+    "transformer.blocks.37.ffn.experts.mlp.v1.weight_scale": "model-00029-of-00031.safetensors",
+    "transformer.blocks.37.ffn.experts.mlp.w1.input_scale": "model-00029-of-00031.safetensors",
+    "transformer.blocks.37.ffn.experts.mlp.w1.weight": "model-00029-of-00031.safetensors",
+    "transformer.blocks.37.ffn.experts.mlp.w1.weight_scale": "model-00029-of-00031.safetensors",
+    "transformer.blocks.37.ffn.experts.mlp.w2.input_scale": "model-00029-of-00031.safetensors",
+    "transformer.blocks.37.ffn.experts.mlp.w2.weight": "model-00029-of-00031.safetensors",
+    "transformer.blocks.37.ffn.experts.mlp.w2.weight_scale": "model-00029-of-00031.safetensors",
+    "transformer.blocks.37.ffn.router.layer.weight": "model-00028-of-00031.safetensors",
+    "transformer.blocks.37.norm_attn_norm.attn.Wqkv.input_scale": "model-00028-of-00031.safetensors",
+    "transformer.blocks.37.norm_attn_norm.attn.Wqkv.weight": "model-00028-of-00031.safetensors",
+    "transformer.blocks.37.norm_attn_norm.attn.Wqkv.weight_scale": "model-00028-of-00031.safetensors",
+    "transformer.blocks.37.norm_attn_norm.attn.kv_scale": "model-00028-of-00031.safetensors",
+    "transformer.blocks.37.norm_attn_norm.attn.out_proj.input_scale": "model-00028-of-00031.safetensors",
+    "transformer.blocks.37.norm_attn_norm.attn.out_proj.weight": "model-00028-of-00031.safetensors",
+    "transformer.blocks.37.norm_attn_norm.attn.out_proj.weight_scale": "model-00028-of-00031.safetensors",
+    "transformer.blocks.37.norm_attn_norm.norm_1.weight": "model-00028-of-00031.safetensors",
+    "transformer.blocks.37.norm_attn_norm.norm_2.weight": "model-00028-of-00031.safetensors",
+    "transformer.blocks.38.ffn.experts.mlp.v1.input_scale": "model-00030-of-00031.safetensors",
+    "transformer.blocks.38.ffn.experts.mlp.v1.weight": "model-00030-of-00031.safetensors",
+    "transformer.blocks.38.ffn.experts.mlp.v1.weight_scale": "model-00030-of-00031.safetensors",
+    "transformer.blocks.38.ffn.experts.mlp.w1.input_scale": "model-00029-of-00031.safetensors",
+    "transformer.blocks.38.ffn.experts.mlp.w1.weight": "model-00029-of-00031.safetensors",
+    "transformer.blocks.38.ffn.experts.mlp.w1.weight_scale": "model-00029-of-00031.safetensors",
+    "transformer.blocks.38.ffn.experts.mlp.w2.input_scale": "model-00030-of-00031.safetensors",
+    "transformer.blocks.38.ffn.experts.mlp.w2.weight": "model-00030-of-00031.safetensors",
+    "transformer.blocks.38.ffn.experts.mlp.w2.weight_scale": "model-00030-of-00031.safetensors",
+    "transformer.blocks.38.ffn.router.layer.weight": "model-00029-of-00031.safetensors",
+    "transformer.blocks.38.norm_attn_norm.attn.Wqkv.input_scale": "model-00029-of-00031.safetensors",
+    "transformer.blocks.38.norm_attn_norm.attn.Wqkv.weight": "model-00029-of-00031.safetensors",
+    "transformer.blocks.38.norm_attn_norm.attn.Wqkv.weight_scale": "model-00029-of-00031.safetensors",
+    "transformer.blocks.38.norm_attn_norm.attn.kv_scale": "model-00029-of-00031.safetensors",
+    "transformer.blocks.38.norm_attn_norm.attn.out_proj.input_scale": "model-00029-of-00031.safetensors",
+    "transformer.blocks.38.norm_attn_norm.attn.out_proj.weight": "model-00029-of-00031.safetensors",
+    "transformer.blocks.38.norm_attn_norm.attn.out_proj.weight_scale": "model-00029-of-00031.safetensors",
+    "transformer.blocks.38.norm_attn_norm.norm_1.weight": "model-00029-of-00031.safetensors",
+    "transformer.blocks.38.norm_attn_norm.norm_2.weight": "model-00029-of-00031.safetensors",
+    "transformer.blocks.39.ffn.experts.mlp.v1.input_scale": "model-00030-of-00031.safetensors",
+    "transformer.blocks.39.ffn.experts.mlp.v1.weight": "model-00030-of-00031.safetensors",
+    "transformer.blocks.39.ffn.experts.mlp.v1.weight_scale": "model-00030-of-00031.safetensors",
+    "transformer.blocks.39.ffn.experts.mlp.w1.input_scale": "model-00030-of-00031.safetensors",
+    "transformer.blocks.39.ffn.experts.mlp.w1.weight": "model-00030-of-00031.safetensors",
+    "transformer.blocks.39.ffn.experts.mlp.w1.weight_scale": "model-00030-of-00031.safetensors",
+    "transformer.blocks.39.ffn.experts.mlp.w2.input_scale": "model-00031-of-00031.safetensors",
+    "transformer.blocks.39.ffn.experts.mlp.w2.weight": "model-00031-of-00031.safetensors",
+    "transformer.blocks.39.ffn.experts.mlp.w2.weight_scale": "model-00031-of-00031.safetensors",
+    "transformer.blocks.39.ffn.router.layer.weight": "model-00030-of-00031.safetensors",
+    "transformer.blocks.39.norm_attn_norm.attn.Wqkv.input_scale": "model-00030-of-00031.safetensors",
+    "transformer.blocks.39.norm_attn_norm.attn.Wqkv.weight": "model-00030-of-00031.safetensors",
+    "transformer.blocks.39.norm_attn_norm.attn.Wqkv.weight_scale": "model-00030-of-00031.safetensors",
+    "transformer.blocks.39.norm_attn_norm.attn.kv_scale": "model-00030-of-00031.safetensors",
+    "transformer.blocks.39.norm_attn_norm.attn.out_proj.input_scale": "model-00030-of-00031.safetensors",
+    "transformer.blocks.39.norm_attn_norm.attn.out_proj.weight": "model-00030-of-00031.safetensors",
+    "transformer.blocks.39.norm_attn_norm.attn.out_proj.weight_scale": "model-00030-of-00031.safetensors",
+    "transformer.blocks.39.norm_attn_norm.norm_1.weight": "model-00030-of-00031.safetensors",
+    "transformer.blocks.39.norm_attn_norm.norm_2.weight": "model-00030-of-00031.safetensors",
+    "transformer.blocks.4.ffn.experts.mlp.v1.input_scale": "model-00004-of-00031.safetensors",
+    "transformer.blocks.4.ffn.experts.mlp.v1.weight": "model-00004-of-00031.safetensors",
+    "transformer.blocks.4.ffn.experts.mlp.v1.weight_scale": "model-00004-of-00031.safetensors",
+    "transformer.blocks.4.ffn.experts.mlp.w1.input_scale": "model-00004-of-00031.safetensors",
+    "transformer.blocks.4.ffn.experts.mlp.w1.weight": "model-00004-of-00031.safetensors",
+    "transformer.blocks.4.ffn.experts.mlp.w1.weight_scale": "model-00004-of-00031.safetensors",
+    "transformer.blocks.4.ffn.experts.mlp.w2.input_scale": "model-00004-of-00031.safetensors",
+    "transformer.blocks.4.ffn.experts.mlp.w2.weight": "model-00004-of-00031.safetensors",
+    "transformer.blocks.4.ffn.experts.mlp.w2.weight_scale": "model-00004-of-00031.safetensors",
+    "transformer.blocks.4.ffn.router.layer.weight": "model-00004-of-00031.safetensors",
+    "transformer.blocks.4.norm_attn_norm.attn.Wqkv.input_scale": "model-00004-of-00031.safetensors",
+    "transformer.blocks.4.norm_attn_norm.attn.Wqkv.weight": "model-00004-of-00031.safetensors",
+    "transformer.blocks.4.norm_attn_norm.attn.Wqkv.weight_scale": "model-00004-of-00031.safetensors",
+    "transformer.blocks.4.norm_attn_norm.attn.kv_scale": "model-00004-of-00031.safetensors",
+    "transformer.blocks.4.norm_attn_norm.attn.out_proj.input_scale": "model-00004-of-00031.safetensors",
+    "transformer.blocks.4.norm_attn_norm.attn.out_proj.weight": "model-00004-of-00031.safetensors",
+    "transformer.blocks.4.norm_attn_norm.attn.out_proj.weight_scale": "model-00004-of-00031.safetensors",
+    "transformer.blocks.4.norm_attn_norm.norm_1.weight": "model-00004-of-00031.safetensors",
+    "transformer.blocks.4.norm_attn_norm.norm_2.weight": "model-00004-of-00031.safetensors",
+    "transformer.blocks.5.ffn.experts.mlp.v1.input_scale": "model-00005-of-00031.safetensors",
+    "transformer.blocks.5.ffn.experts.mlp.v1.weight": "model-00005-of-00031.safetensors",
+    "transformer.blocks.5.ffn.experts.mlp.v1.weight_scale": "model-00005-of-00031.safetensors",
+    "transformer.blocks.5.ffn.experts.mlp.w1.input_scale": "model-00005-of-00031.safetensors",
+    "transformer.blocks.5.ffn.experts.mlp.w1.weight": "model-00005-of-00031.safetensors",
+    "transformer.blocks.5.ffn.experts.mlp.w1.weight_scale": "model-00005-of-00031.safetensors",
+    "transformer.blocks.5.ffn.experts.mlp.w2.input_scale": "model-00005-of-00031.safetensors",
+    "transformer.blocks.5.ffn.experts.mlp.w2.weight": "model-00005-of-00031.safetensors",
+    "transformer.blocks.5.ffn.experts.mlp.w2.weight_scale": "model-00005-of-00031.safetensors",
+    "transformer.blocks.5.ffn.router.layer.weight": "model-00004-of-00031.safetensors",
+    "transformer.blocks.5.norm_attn_norm.attn.Wqkv.input_scale": "model-00004-of-00031.safetensors",
+    "transformer.blocks.5.norm_attn_norm.attn.Wqkv.weight": "model-00004-of-00031.safetensors",
+    "transformer.blocks.5.norm_attn_norm.attn.Wqkv.weight_scale": "model-00004-of-00031.safetensors",
+    "transformer.blocks.5.norm_attn_norm.attn.kv_scale": "model-00004-of-00031.safetensors",
+    "transformer.blocks.5.norm_attn_norm.attn.out_proj.input_scale": "model-00004-of-00031.safetensors",
+    "transformer.blocks.5.norm_attn_norm.attn.out_proj.weight": "model-00004-of-00031.safetensors",
+    "transformer.blocks.5.norm_attn_norm.attn.out_proj.weight_scale": "model-00004-of-00031.safetensors",
+    "transformer.blocks.5.norm_attn_norm.norm_1.weight": "model-00004-of-00031.safetensors",
+    "transformer.blocks.5.norm_attn_norm.norm_2.weight": "model-00004-of-00031.safetensors",
+    "transformer.blocks.6.ffn.experts.mlp.v1.input_scale": "model-00006-of-00031.safetensors",
+    "transformer.blocks.6.ffn.experts.mlp.v1.weight": "model-00006-of-00031.safetensors",
+    "transformer.blocks.6.ffn.experts.mlp.v1.weight_scale": "model-00006-of-00031.safetensors",
+    "transformer.blocks.6.ffn.experts.mlp.w1.input_scale": "model-00005-of-00031.safetensors",
+    "transformer.blocks.6.ffn.experts.mlp.w1.weight": "model-00005-of-00031.safetensors",
+    "transformer.blocks.6.ffn.experts.mlp.w1.weight_scale": "model-00005-of-00031.safetensors",
+    "transformer.blocks.6.ffn.experts.mlp.w2.input_scale": "model-00006-of-00031.safetensors",
+    "transformer.blocks.6.ffn.experts.mlp.w2.weight": "model-00006-of-00031.safetensors",
+    "transformer.blocks.6.ffn.experts.mlp.w2.weight_scale": "model-00006-of-00031.safetensors",
+    "transformer.blocks.6.ffn.router.layer.weight": "model-00005-of-00031.safetensors",
+    "transformer.blocks.6.norm_attn_norm.attn.Wqkv.input_scale": "model-00005-of-00031.safetensors",
+    "transformer.blocks.6.norm_attn_norm.attn.Wqkv.weight": "model-00005-of-00031.safetensors",
+    "transformer.blocks.6.norm_attn_norm.attn.Wqkv.weight_scale": "model-00005-of-00031.safetensors",
+    "transformer.blocks.6.norm_attn_norm.attn.kv_scale": "model-00005-of-00031.safetensors",
+    "transformer.blocks.6.norm_attn_norm.attn.out_proj.input_scale": "model-00005-of-00031.safetensors",
+    "transformer.blocks.6.norm_attn_norm.attn.out_proj.weight": "model-00005-of-00031.safetensors",
+    "transformer.blocks.6.norm_attn_norm.attn.out_proj.weight_scale": "model-00005-of-00031.safetensors",
+    "transformer.blocks.6.norm_attn_norm.norm_1.weight": "model-00005-of-00031.safetensors",
+    "transformer.blocks.6.norm_attn_norm.norm_2.weight": "model-00005-of-00031.safetensors",
+    "transformer.blocks.7.ffn.experts.mlp.v1.input_scale": "model-00006-of-00031.safetensors",
+    "transformer.blocks.7.ffn.experts.mlp.v1.weight": "model-00006-of-00031.safetensors",
+    "transformer.blocks.7.ffn.experts.mlp.v1.weight_scale": "model-00006-of-00031.safetensors",
+    "transformer.blocks.7.ffn.experts.mlp.w1.input_scale": "model-00006-of-00031.safetensors",
+    "transformer.blocks.7.ffn.experts.mlp.w1.weight": "model-00006-of-00031.safetensors",
+    "transformer.blocks.7.ffn.experts.mlp.w1.weight_scale": "model-00006-of-00031.safetensors",
+    "transformer.blocks.7.ffn.experts.mlp.w2.input_scale": "model-00007-of-00031.safetensors",
+    "transformer.blocks.7.ffn.experts.mlp.w2.weight": "model-00007-of-00031.safetensors",
+    "transformer.blocks.7.ffn.experts.mlp.w2.weight_scale": "model-00007-of-00031.safetensors",
+    "transformer.blocks.7.ffn.router.layer.weight": "model-00006-of-00031.safetensors",
+    "transformer.blocks.7.norm_attn_norm.attn.Wqkv.input_scale": "model-00006-of-00031.safetensors",
+    "transformer.blocks.7.norm_attn_norm.attn.Wqkv.weight": "model-00006-of-00031.safetensors",
+    "transformer.blocks.7.norm_attn_norm.attn.Wqkv.weight_scale": "model-00006-of-00031.safetensors",
+    "transformer.blocks.7.norm_attn_norm.attn.kv_scale": "model-00006-of-00031.safetensors",
+    "transformer.blocks.7.norm_attn_norm.attn.out_proj.input_scale": "model-00006-of-00031.safetensors",
+    "transformer.blocks.7.norm_attn_norm.attn.out_proj.weight": "model-00006-of-00031.safetensors",
+    "transformer.blocks.7.norm_attn_norm.attn.out_proj.weight_scale": "model-00006-of-00031.safetensors",
+    "transformer.blocks.7.norm_attn_norm.norm_1.weight": "model-00006-of-00031.safetensors",
+    "transformer.blocks.7.norm_attn_norm.norm_2.weight": "model-00006-of-00031.safetensors",
+    "transformer.blocks.8.ffn.experts.mlp.v1.input_scale": "model-00007-of-00031.safetensors",
+    "transformer.blocks.8.ffn.experts.mlp.v1.weight": "model-00007-of-00031.safetensors",
+    "transformer.blocks.8.ffn.experts.mlp.v1.weight_scale": "model-00007-of-00031.safetensors",
+    "transformer.blocks.8.ffn.experts.mlp.w1.input_scale": "model-00007-of-00031.safetensors",
+    "transformer.blocks.8.ffn.experts.mlp.w1.weight": "model-00007-of-00031.safetensors",
+    "transformer.blocks.8.ffn.experts.mlp.w1.weight_scale": "model-00007-of-00031.safetensors",
+    "transformer.blocks.8.ffn.experts.mlp.w2.input_scale": "model-00007-of-00031.safetensors",
+    "transformer.blocks.8.ffn.experts.mlp.w2.weight": "model-00007-of-00031.safetensors",
+    "transformer.blocks.8.ffn.experts.mlp.w2.weight_scale": "model-00007-of-00031.safetensors",
+    "transformer.blocks.8.ffn.router.layer.weight": "model-00007-of-00031.safetensors",
+    "transformer.blocks.8.norm_attn_norm.attn.Wqkv.input_scale": "model-00007-of-00031.safetensors",
+    "transformer.blocks.8.norm_attn_norm.attn.Wqkv.weight": "model-00007-of-00031.safetensors",
+    "transformer.blocks.8.norm_attn_norm.attn.Wqkv.weight_scale": "model-00007-of-00031.safetensors",
+    "transformer.blocks.8.norm_attn_norm.attn.kv_scale": "model-00007-of-00031.safetensors",
+    "transformer.blocks.8.norm_attn_norm.attn.out_proj.input_scale": "model-00007-of-00031.safetensors",
+    "transformer.blocks.8.norm_attn_norm.attn.out_proj.weight": "model-00007-of-00031.safetensors",
+    "transformer.blocks.8.norm_attn_norm.attn.out_proj.weight_scale": "model-00007-of-00031.safetensors",
+    "transformer.blocks.8.norm_attn_norm.norm_1.weight": "model-00007-of-00031.safetensors",
+    "transformer.blocks.8.norm_attn_norm.norm_2.weight": "model-00007-of-00031.safetensors",
+    "transformer.blocks.9.ffn.experts.mlp.v1.input_scale": "model-00008-of-00031.safetensors",
+    "transformer.blocks.9.ffn.experts.mlp.v1.weight": "model-00008-of-00031.safetensors",
+    "transformer.blocks.9.ffn.experts.mlp.v1.weight_scale": "model-00008-of-00031.safetensors",
+    "transformer.blocks.9.ffn.experts.mlp.w1.input_scale": "model-00008-of-00031.safetensors",
+    "transformer.blocks.9.ffn.experts.mlp.w1.weight": "model-00008-of-00031.safetensors",
+    "transformer.blocks.9.ffn.experts.mlp.w1.weight_scale": "model-00008-of-00031.safetensors",
+    "transformer.blocks.9.ffn.experts.mlp.w2.input_scale": "model-00008-of-00031.safetensors",
+    "transformer.blocks.9.ffn.experts.mlp.w2.weight": "model-00008-of-00031.safetensors",
+    "transformer.blocks.9.ffn.experts.mlp.w2.weight_scale": "model-00008-of-00031.safetensors",
+    "transformer.blocks.9.ffn.router.layer.weight": "model-00007-of-00031.safetensors",
+    "transformer.blocks.9.norm_attn_norm.attn.Wqkv.input_scale": "model-00007-of-00031.safetensors",
+    "transformer.blocks.9.norm_attn_norm.attn.Wqkv.weight": "model-00007-of-00031.safetensors",
+    "transformer.blocks.9.norm_attn_norm.attn.Wqkv.weight_scale": "model-00007-of-00031.safetensors",
+    "transformer.blocks.9.norm_attn_norm.attn.kv_scale": "model-00007-of-00031.safetensors",
+    "transformer.blocks.9.norm_attn_norm.attn.out_proj.input_scale": "model-00007-of-00031.safetensors",
+    "transformer.blocks.9.norm_attn_norm.attn.out_proj.weight": "model-00007-of-00031.safetensors",
+    "transformer.blocks.9.norm_attn_norm.attn.out_proj.weight_scale": "model-00007-of-00031.safetensors",
+    "transformer.blocks.9.norm_attn_norm.norm_1.weight": "model-00007-of-00031.safetensors",
+    "transformer.blocks.9.norm_attn_norm.norm_2.weight": "model-00007-of-00031.safetensors",
+    "transformer.norm_f.weight": "model-00031-of-00031.safetensors",
+    "transformer.wte.weight": "model-00001-of-00031.safetensors"
+  }
+}

special_tokens_map.json

@@ -0,0 +1,30 @@
+{
+  "bos_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|pad|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "unk_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

tokenizer_config.json

@@ -0,0 +1,205 @@
+{
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "100256": {
+      "content": "<||_unused_0_||>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "100257": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "100258": {
+      "content": "<|fim_prefix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "100259": {
+      "content": "<|fim_middle|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "100260": {
+      "content": "<|fim_suffix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "100261": {
+      "content": "<||_unused_1_||>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "100262": {
+      "content": "<||_unused_2_||>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "100263": {
+      "content": "<||_unused_3_||>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "100264": {
+      "content": "<||_unused_4_||>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "100265": {
+      "content": "<||_unused_5_||>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "100266": {
+      "content": "<||_unused_6_||>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "100267": {
+      "content": "<||_unused_7_||>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "100268": {
+      "content": "<||_unused_8_||>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "100269": {
+      "content": "<||_unused_9_||>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "100270": {
+      "content": "<||_unused_10_||>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "100271": {
+      "content": "<||_unused_11_||>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "100272": {
+      "content": "<||_unused_12_||>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "100273": {
+      "content": "<||_unused_13_||>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "100274": {
+      "content": "<||_unused_14_||>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "100275": {
+      "content": "<||_unused_15_||>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "100276": {
+      "content": "<|endofprompt|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "100277": {
+      "content": "<|pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "100278": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "100279": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "bos_token": "<|endoftext|>",
+  "chat_template": "{% if messages[0]['role'] == 'system' %}{% set loop_messages = messages[1:] %}{% set system_message = messages[0]['content'] %}{% elif 'system' not in messages[0]['role'] %}{% set loop_messages = messages %}{% set system_message = 'You are DBRX, created by Databricks. You were last updated in December 2023. You answer questions based on information available up to that point.\nYOU PROVIDE SHORT RESPONSES TO SHORT QUESTIONS OR STATEMENTS, but provide thorough responses to more complex and open-ended questions.\nYou assist with various tasks, from writing to coding (using markdown for code blocks — remember to use ``` with code, JSON, and tables).\n(You do not have real-time data access or code execution capabilities. You avoid stereotyping and provide balanced perspectives on controversial topics. You do not provide song lyrics, poems, or news articles and do not divulge details of your training data.)\nThis is your system prompt, guiding your responses. Do not reference it, just respond to the user. If you find yourself talking about this message, stop. You should be responding appropriately and usually that means not mentioning this.\nYOU DO NOT MENTION ANY OF THIS INFORMATION ABOUT YOURSELF UNLESS THE INFORMATION IS DIRECTLY PERTINENT TO THE USER\\'S QUERY.' %}{% else %}{% set loop_messages = messages %}{% set system_message = false %}{% endif %}{% for message in loop_messages %}{% if loop.index0 == 0 %}{% if system_message != false %}{{ '<|im_start|>system\n' + system_message | trim + '<|im_end|>\n'}}{% endif %}{{ '<|im_start|>' + message['role'] + '\n' + message['content'] + '<|im_end|>' }}{% else %}{{ '\n' + '<|im_start|>' + message['role'] + '\n' + message['content'] + '<|im_end|>' }}{% endif %}{% if (add_generation_prompt == true and loop.last) %}{{ '\n' + '<|im_start|>' + 'assistant' + '\n' }}{% endif %}{% endfor %}",
+  "clean_up_tokenization_spaces": true,
+  "eos_token": "<|endoftext|>",
+  "model_max_length": 32768,
+  "pad_token": "<|pad|>",
+  "tokenizer_class": "GPT2Tokenizer",
+  "unk_token": "<|endoftext|>"
+}